This invention relates to a self healing system for restoring a communication network assigned with cross-connect functions from various failures and, more particularly, for currently used paths, each between a source and a destination node of communication.
(Prior Art)
Such a self healing system has hitherto been devised or used in order to restore communication services from network failures caused by faults in links or nodes.
Upon occurrence of a failure in a currently used path, a restoration path must be established in the network in order to avoid a break in communication. For this self heal, the network is preliminarily assigned with capacities and like reserved resources. The reserved resources are controlled for self heal of a failure, which may take place, according to either a centralized control scheme or a distributed control scheme.
According to the centralized control scheme, failure situations of the network are transmitted to a specific central control node. Based on such accumulated information, calculated is a route of an optimum restoration path, namely, a self heal countermeasure.
According to the distributed control scheme, no central control node is used. Instead, control signals are sent and received at nodes of the network to autonomously and distributedly search for a restoration path. Reported is a restoration performance composed of various numbers of phases. It will be assumed here that two phases are used. In a first phase (a broadcast phase), a sender (a first terminal node which is determined in the currently used path prior to occurrence of a failure) detects a failure and sends a restoration signal to all adjacent links which satisfy restoration requirements. This operation is repeated until a chooser (a second terminal node of the currently used path subjected to the failure) receives the restoration signal. In a second phase (a connection phase), the chooser selects a restoration path from tested paths and sends a connection signal through a decided route. This is in order to establish the restoration path. In this manner, the restoration signal and the connection signal are twice sent and received at both path terminal nodes to complete self heal of the path.
In contrast, a self healing system is proposed in a paper contributed by Ryutaro Kawamura, Ken-ichi Sato, and Ikuo Tokizawa to the Proceedings of 5th International Network Planning Symposium (Networks '92), pages 129 to 134, May 1992, under the title of "Self Healing ATM Network Techniques Utilizing Virtual Paths", wherein two algorithms are proposed to shorten a time of restoration. In an algorithm 1 (Flood-Connecting algorithm), the broadcast phase and the connection phase are degenerated into only one phase. In another algorithm 2 (Pre-assigned backup VP algorithm), a restoration path is preassigned to the currently used path to send a message only to the restoration path when a failure takes place. Should a multiple failure occur so as to render the restoration path unavailable, the failure is tried for restoration in accordance with the algorithm 1 between the terminal nodes of the currently used path.
(Tasks to Be Solved by the Invention)
With the conventional centralized control method, it is possible to optimize the self heal based on information of the network as a whole and to economize the self heal. It, however, takes time to collect the information at the central control node, to calculate the countermeasure for the self heal, and to transmit switching information to all nodes. This gives rise to a task of unavoidance of an appreciably long self heal time.
With the algorithm 1 which is an improvement in the conventional distributed control scheme, it is possible to achieve a short self heal time. It is, however, indispensable to transfer control signals to all adjacent nodes which are available. This results in transfer of an enormous number of control signals in the network for the self heal. It is unavoidable to temporarily establish a plurality of restoration paths for one currently used path to use link resources in vain. This may result in no serious problem for an isolated failure. This, however, results in the multiple failure in congestion of the control signals for the self heal and in deterioration of performance of a self heal rate and of the restoration time to give rise to an instability reported in a paper contributed by Hiroyuki Fujii and Noriaki Yoshikai to the Proceedings of NOMS 92, pages 231 to 241, 1992, under the title of "Transfer Mechanism and Operation of Self-Healing Algorithms in ATM Networks". Alternatively, this results in a task of complexed control because the multiple failure must be traded off by assignment of a priority degree to messages for the self heal.
With the algorithm 2 which is another improvement in the conventional distributed control scheme, restoration is scheduled in the multiple failure according to the algorithm 1. As a consequence, a task is involved in the case of the multiple failure.
It is consequently an object of the present invention to provide a novel self healing system having merits of the centralized control scheme and the distributed control scheme to attain advantages of these both schemes and to remove the above-mentioned defects.